1. Field of the Invention
The present invention relates to a mobile communications system and, more particularly, to a method for multiplexing control signals and reference signals (also referred to as pilot signals), a method for allocating resources, and a base station using the resource allocation method.
2. Description of the Related Art
In the Third Generation Partnership Project (3GPP), standardization of Long Term Evolution (LTE), so-called 3.9G, is currently progressing. In LTE, single-carrier transmission is considered as an uplink access scheme. It can be said that the single-carrier transmission is an access scheme excellent for power efficiency in comparison with multi-carrier transmission, such as orthogonal frequency division multiplexing (OFDM), because the peak-to-average power ratio (PAPR) can be kept low. Hence, it can be said that the single-carrier transmission is an access scheme suitable for uplink.
FIG. 1A is a diagram showing a frame format for uplink supported by LTE, which is described in 3GPP, “TR 25.814 v7.0.0,” Section 9.1.1. In LTE, communication is performed in units of a frame (also referred to as a sub-frame) of a time length of 0.5 msec. One frame includes six long blocks LB#1 to LB#6 and two short blocks SB#1 and SB#2, with a cyclic prefix (CP) added to each block, which will be described later. The time length of a long block is set to be twice as long as that of a short block, and the number of subcarriers in a long block is set to be twice as large as that in a short block. In addition, a subcarrier interval in a long block is set to be half a subcarrier interval in a short block.
Note that although two short blocks are provided here, the number of short blocks, which are allocated for reference signals, depends on the length of a frame, an allowable overhead, and the like. Moreover, as for the timings of the short blocks SB#1 and SB#2 in a frame, the structure shown in FIG. 1A is not limitative, and it suffices to determine the timings so as to allow the reference signals to function effectively.
CAZAC (Constant Amplitude Zero Auto-Correlation) sequence is a predominant one of the sequences used for uplink reference signals. For example, Zadoff-Chu sequence is one type of the CAZAC sequence, represented by the following equation (see Popvic, B. M., “Generalized Chirp-Like Polyphase Sequences with Optimum Correlation Properties,” IEEE Transactions on Information Theory, Vol. 38, No. 4 (July 1992), pp. 1406-1409):
            c      k        ⁡          (      n      )        =      {                                        exp            ⁡                          [                                                                    j                    ⁢                                                                                  ⁢                    2                    ⁢                    π                    ⁢                                                                                  ⁢                    k                                    L                                ⁢                                  (                                                                                    n                        2                                            2                                        +                    n                                    )                                            ]                                                            when            ⁢                                                  ⁢            the            ⁢                                                  ⁢            sequence            ⁢                                                  ⁢            length            ⁢                                                  ⁢            L                                                                                                            is            ⁢                                                  ⁢            an            ⁢                                                  ⁢            even            ⁢                                                  ⁢            number                                                            exp            ⁡                          [                                                                    j                    ⁢                                                                                  ⁢                    2                    ⁢                    π                    ⁢                                                                                  ⁢                    k                                    L                                ⁢                                  (                                                            n                      ⁢                                                                        n                          +                          1                                                2                                                              +                    n                                    )                                            ]                                                            when            ⁢                                                  ⁢            the            ⁢                                                  ⁢            sequence            ⁢                                                  ⁢            length                                                                                                            L            ⁢                                                  ⁢            is            ⁢                                                  ⁢            an            ⁢                                                  ⁢            odd            ⁢                                                  ⁢            number                              where n=0, 1, . . . , and L−1, and k is a sequence number, which is an integer prime to L.
The CAZAC sequence is a sequence that makes the amplitude of a signal constant in time and frequency domains and that allows the autocorrelation value to be zero at a phase difference of any value other than zero. Because of the constant amplitude in time domain, PAPR can be kept low, and because of the constant amplitude in frequency domain as well, the CAZAC sequence is suitable for channel estimation in frequency domain. Moreover, because of the property of perfect autocorrelation, the CAZAC sequence also has the advantage of being suitable to detect the timing of a received signal. For these reasons, the CAZAC sequence has been attracting attention as a sequence suitable for single-carrier transmission. However, in the case of the CAZAC sequence, there is a limit to the number of sequences that can be obtained. The number of sequences depends on the sequence length. In the case of the Zadoff-Chu sequence, the number of sequences reaches it peak when the sequence length L is a prime number, and the maximum number of sequences is equivalent to (L−1).
In uplink, it is necessary that each mobile station (hereinafter, also expressed as UE) transmit a reference signal. Therefore, a variety of methods for multiplexing reference signals of multiple UEs have been proposed.
In 3GPP, R1-051062, Texas Instruments, “On Uplink Pilot in EUTRA SC-FDMA,” October 2005, code division multiplexing (CDM) is proposed as a multiplexing method employed when the CAZAC sequence is used for uplink reference signals.
FIG. 1B is a schematic diagram for describing a method for allotting a CAZAC sequence to a reference signal of each UE. In code-division-multiplexing of reference signals, UEs use CAZAC sequences of the same length, and each UE is assigned a CAZAC sequence having a unique cyclic prefix added thereto as shown in FIG. 1B. If the time length of this cyclic prefix is set to be not shorter than a maximum delay time supposed, then the reference signals of all the UEs can be orthogonalized even in multi-path environments. This is because the autocorrelation value of a CAZAC sequence is always zero except when the phase difference is zero. Note, however, that there is a limit on the number of UEs that can be multiplexed by CDM with respect to reference signal. In a current LTE system, the number of UEs that can be multiplexed is six or so (see 3GPP, R1-060388, Motorola, “Performance Comparison of Pilot/Reference Signal Structures for E-UTRA Uplink SC-FDMA,” February 2006).
An uplink control signal can be classified as any one of a data-dependent control signal (also referred to as a data-associated control signal), which is a control signal regarding uplink data, and a data-independent control signal (also referred to as a data-non-associated control signal), which is feedback information regarding a downlink signal. The data-dependent control signal is a signal transmitted when uplink data is present. If a data-dependent control signal is transmitted by using a resource (long block LB) for transmitting a data signal, an essentially required reference signal (transmitted by using a short block SB) for demodulating the data signal can also be utilized to demodulate the data-dependent control signal.
On the other hand, the data-independent control signal is a signal transmitted as a feedback on downlink data, or the like, and is a signal transmitted independently of an uplink data signal. Accordingly, a reference signal for demodulating the data-independent control signal is required, and the problem of how to allocate a resource for the reference signal arises.
In the foregoing, data-dependent and data-independent control signals have been described with respect to uplink control signal. However, with respect to downlink control signal as well, it can be said that a control signal transmitted when downlink data is present (a data-dependent control signal, which is a control signal regarding downlink data) is a downlink data-dependent control signal, and that a signal transmitted independently of a downlink data signal (a data-independent control signal, which is feedback information regarding an uplink signal) is a downlink data-independent control signal. Hereinafter, to simplify expression, it is assumed that a “control signal” indicates a “data-independent control signal.”
Examples of control information, which is contained in an uplink data-independent control signal, at least include Acknowledgment/Negative Acknowledgment (hereinafter, expressed as Ack/Nack) indicating whether or not downlink information has been received without errors, channel quality indication information (channel quality indicator: hereinafter, expressed as CQI) indicating the state of a downlink channel, a combination of these, and the like. It is desirable that Ack/Nack be transmitted at every transmission time interval (hereinafter, expressed as TTI). However, with the transmission overhead being considered, it is not always necessary to transmit CQI at every TTI. For this reason, there are some occasions when the frequency of transmission of Ack/Nack differs from the frequency of transmission of CQI. Accordingly, within a TTI, UEs transmitting three types of control signals may coexist: a UE transmitting Ack/Nack only, a UE transmitting CQI only, and a UE transmitting both of Ack/Nack and CQI. Incidentally, TTI is a time interval equivalent to a set of multiple blocks (also referred to as a transport block set) transported at a time between the physical and MAC layers.
However, the amount of information of Ack/Nack is smaller than that of CQI. It is possible to make transmission bandwidths of the above-mentioned three types of control signals constant by changing the rate of encoding, or transmitting dummy bits. However, if these transmission bandwidths are made constant, waste occurs with a resource (transmission bandwidth) used to transmit a signal having a small amount of information. To avoid the occurrence of such waste of resource, frequency resources (transmission bandwidths) allocated to transmit the respective control signals are, in general, different transmission bandwidths of three types.
In addition, the transmission made by a UE simultaneously transmitting Ack/Nack and CQI is multi-carrier transmission if corresponding control resources are mapped in uncontiguous frequency bands, resulting in increased PAPR. Accordingly, in order for a UE simultaneously transmitting Ack/Nack and CQI to make single-carrier transmission, resources in adjacent frequency bands need to be allocated to the UE, and these signals need to be processed together. This will be described more specifically with reference to FIGS. 2A and 2B.
FIG. 2A is a diagram showing frequency resource allocation in the case of multi-carrier transmission of control signals, and FIG. 2B is a diagram showing frequency resource allocation in the case of single-carrier transmission of control signals. Referring to FIG. 2A, when control resources in uncontiguous frequency bands F1 and F2 are allocated to a UE which simultaneously transmits Ack/Nack and CQI, the UE cannot perform single-carrier transmission. Accordingly, PAPR is increased as described above.
Therefore, the frequency resources for transmitting Ack/Nack and CQI are mapped into a combined band of adjacent frequency bands F3 and F4 as shown in FIG. 2B, whereby these bands can be handled as a single band, enabling single-carrier transmission.
FIG. 3 is a diagram showing an example of the allocation of resources for control and reference signals. Here, shown is the case, as an example, where control signals in a long block LB#1 and reference signals in a short block SB#1 are time-division-multiplexed (TDM). Incidentally, a numeral applied to each control or reference signal in the drawing represents a UE's number (the same goes for the other drawings.)
As to the control signals regarding downlink data signals, there are three types of UEs coexisting, each transmitting Ack/Nack only, CQI only, or both of Ack/Nack and CQI, as described above. Here, UEs 1 and 6 each transmit both of Ack/Nack and CQI, UEs 2 and 3 each transmit Ack/Nack only, and UEs 4 and 5 each transmit CQI only.
However, according to a conventional resource allocation method, a reference signal is allocated a reference resource in the same bandwidth which a control signal to be demodulated is transmitted in. That is, reference resources each corresponding to three types of transmission bandwidths are to be allocated. Therefore, a reference signal for demodulating Ack/Nack, which has a small amount of information and hence a small transmission bandwidth, also has a reduced transmission bandwidth. Since the length of the CAZAC sequence, which is used for reference signals, depends on the transmission bandwidth as described above, the number of usable reference signal sequences (CAZAC sequences) decreases when Ack/Nack only is transmitted.
However, the number of reference signal sequences is an important factor to the cell designing in a cellular system composed of multiple cells. The reason is that the use of the same reference signal sequence by adjacent cells leads to increased interference between the cells, and to avoid this, adjacent cells need to use different reference signal sequences. According to the conventional resource allocation method, as described above, if a transmission bandwidth is small as in the case of Ack/Nack, the length of the usable CAZAC sequence is short. Therefore, the problem arises that there occurs a shortage of the sequences to be used for reference signals at the time of transmission.